@MartinBüttner any ideas for this for the palindrome question? Feels kind of fishy in a couple places. Particularly the leading space thing, is there a better way to get the prefixes?

Also, I did try all the test cases, but it does also seem sort of weird that this works? Do you see any cases where it'd mess up?

Actually a lot of that is unnecessary: retina.tryitonline.net/…

retina.tryitonline.net/…

retina.tryitonline.net/…

retina.tryitonline.net/…

Ahh so many golfs.

Also I didn't know that about the numerical sort, that's smart.

if \d+ doesn't match the string, it defaults to 0

Thanks for all the help :) I think it looks pretty good now, I guess I'll post it.

reversing the first string is 4 bytes cheaper, but I don't see an easy way to swap the strings afterwards

Hmm, yeah O#`\w* is too long...

Yeah, that'd be useful, maybe like $-<group>?

I was thinking $< or $R but yeah, something

actually not <group>

could be any token... so $<$& would reverse the whole match

but maybe you're right and doing something like I did with $. and $# where the flag is inserted to a normal substitution element would be even better

Hrm, yeah I don't know when it would be better with the token way. Unless maybe a lot of fancy new token things are on the horizon :P

I'm more worried about how combining those flags would work :P

Ah, right that makes sense :P

So like $<#1 would count the number of matches in the first group and then reverse it kind of thing?

yeah, it would reverse the decimal... not sure how useful that is, but why not...

Yeah probably not super useful right now, but with more flags it probably gets better :P

Also, how did you get reversing the first one to be 4 bytes shorter? I can't figure it out >_<

\G\w

Ohhh that's smart.

\G. if you use ¶ as the separator

oh wow

retina.tryitonline.net/…

I didn't know \G worked in a useful way with RTL mode

Oh wow, that is neat